def test_NN2_params_rand2(self): h2o.beta_features = True csvPathname = 'covtype/covtype.20k.data' hex_key = 'covtype.20k.hex' parseResult = h2i.import_parse(bucket='smalldata', path=csvPathname, hex_key=hex_key, schema='put') paramDict = define_params() for trial in range(3): # params is mutable. This is default. params = {'response': 'C55', 'epochs': '1'} h2o_nn.pickRandDeepLearningParams(paramDict, params) kwargs = params.copy() start = time.time() nn = h2o_cmd.runDeepLearning(timeoutSecs=500, parseResult=parseResult, **kwargs) print "nn result:", h2o.dump_json(nn) h2o.check_sandbox_for_errors() # FIX! simple check? print "Deep Learning end on ", csvPathname, 'took', time.time( ) - start, 'seconds' print "Trial #", trial, "completed\n"
def test_NN2_params_rand2(self): csvPathname = 'covtype/covtype.20k.data' hex_key = 'covtype.20k.hex' parseResult = h2i.import_parse(bucket='smalldata', path=csvPathname, hex_key=hex_key, schema='put') paramDict = define_params() for trial in range(5): # params is mutable. This is default. params = {'response': 'C55'} h2o_nn.pickRandDeepLearningParams(paramDict, params) kwargs = params.copy() start = time.time() nn = h2o_cmd.runDeepLearning(timeoutSecs=300, parseResult=parseResult, **kwargs) print "nn result:", h2o.dump_json(nn) h2o.check_sandbox_for_errors() # FIX! simple check? print "Deep Learning end on ", csvPathname, 'took', time.time() - start, 'seconds' print "Trial #", trial, "completed\n"
def test_NN2_params_rand2(self): csvPathname = 'covtype/covtype.20k.data' hex_key = 'covtype.20k.hex' parseResult = h2i.import_parse(bucket='smalldata', path=csvPathname, hex_key=hex_key, schema='put') paramDict = define_params() for trial in range(3): # params is mutable. This is default. params = {'response': 'C55', 'epochs': '1'} h2o_nn.pickRandDeepLearningParams(paramDict, params) kwargs = params.copy() start = time.time() nn = h2o_cmd.runDeepLearning(timeoutSecs=500, parseResult=parseResult, **kwargs) print "nn result:", h2o.dump_json(nn) h2o.check_sandbox_for_errors() deeplearning_model = nn['deeplearning_model'] errors = deeplearning_model['errors'] # print "errors", h2o.dump_json(errors) # print "errors, classification", errors['classification'] # assert 1==0 # unstable = nn['model_info']['unstable'] # unstable case caused by : # normal initial distribution with amplitude 1 and input_dropout_ratio=1. # blowing up numerically during propagation of all zeroes as input repeatedly. # arnon added logging to stdout in addition to html in 7899b92ad67. # Will have to check that first before making predictions. # print "unstable:", unstable # FIX! simple check? print "Deep Learning end on ", csvPathname, 'took', time.time( ) - start, 'seconds' print "Trial #", trial, "completed\n"
def test_NN2_params_rand2(self): h2o.beta_features = True csvPathname = 'covtype/covtype.20k.data' hex_key = 'covtype.20k.hex' parseResult = h2i.import_parse(bucket='smalldata', path=csvPathname, hex_key=hex_key, schema='put') paramDict = define_params() for trial in range(3): # params is mutable. This is default. params = {'response': 'C55', 'epochs': '1'} h2o_nn.pickRandDeepLearningParams(paramDict, params) kwargs = params.copy() start = time.time() nn = h2o_cmd.runDeepLearning(timeoutSecs=500, parseResult=parseResult, **kwargs) print "nn result:", h2o.dump_json(nn) h2o.check_sandbox_for_errors() deeplearning_model = nn['deeplearning_model'] errors = deeplearning_model['errors'] # print "errors", h2o.dump_json(errors) # print "errors, classification", errors['classification'] # assert 1==0 # unstable = nn['model_info']['unstable'] # unstable case caused by : # normal initial distribution with amplitude 1 and input_dropout_ratio=1. # blowing up numerically during propagation of all zeroes as input repeatedly. # arnon added logging to stdout in addition to html in 7899b92ad67. # Will have to check that first before making predictions. # print "unstable:", unstable # FIX! simple check? print "Deep Learning end on ", csvPathname, 'took', time.time() - start, 'seconds' print "Trial #", trial, "completed\n"